Abstract
Many video games, especially those developed by indie companies, are translated by volunteers without training in translation. The consequences of this practice are yet to be studied in depth and this article will try to contribute to the research on the differences in the localized games depending on the type of translation provider. Through a survey of the developers of 172 video games, some details such as the profile of the translator or the revision level applied have been confirmed. Thanks to this information, we have built a multimodal corpus of four games translated by professionals (with higher or lower revision and quality control levels), and two games translated by volunteers. The preliminary data and the macrostructural study of the corpus have revealed the context in which the videogames were localized and that no changes were applied to elements outside the written and spoken discourse (there were no changes to the background music or graphic icons, regardless of the type of translator). The microstructural analysis, in turn, has highlighted linguistic differences among games, being those translated by volunteers the ones that include more mistakes. In the analysed corpus, the subgroup of videogames translated by volunteers presents up to seven times more phenomena noticeable by the end player than the subgroup translated by professionals with the more exhaustive localisation process.
Keywords
video games; localization; audiovisual translation; multimodal corpus; volunteers